What Is Audio Latency? Causes and How to Reduce It

Audio latency is the delay between a sound or action and when you hear or record its result. If you speak into a microphone and hear your voice noticeably late in headphones, the signal may be traveling through an interface, computer, recording software and back to the output. Buffering and processing each contribute time.

Some delay is normal. The goal is a monitoring setup that feels usable and records reliably, not the smallest number at any cost. If lowering a buffer causes clicks or dropouts, raise it until the system is stable.

Where does audio latency come from?

Part of the path	What can add delay	A useful check
Interface and driver	Conversion, hardware and driver buffering.	Select the intended interface and use the driver recommended by its manufacturer.
Recording app	Input and output buffers hold audio in chunks so the computer can process it consistently.	Lower the buffer one step, then listen and record a test for dropouts.
Effects	Some plug-ins need extra processing time, including effects with look-ahead.	Temporarily bypass suspect effects and compare while monitoring.
Wireless output	Bluetooth headphones or speakers can add delay.	Compare wired headphones connected to the appropriate output.
Remote call	Network transport and other participants’ devices add their own delays.	Test locally first; an interface buffer change cannot eliminate internet delay.

Apple’s Logic Pro guide lists buffers, conversion, interface software, sample rate, plug-ins and Bluetooth among factors in input monitoring latency. Ableton’s explanation also traces latency through the interface, driver, operating system and DAW.

Buffer size and sample rate: an example

A buffer stores a short block of samples. Its approximate duration in milliseconds is buffer size in samples ÷ sample rate in kHz. At 48 kHz, a 128-sample buffer represents about 2.67 ms for one buffer; 256 samples represent about 5.33 ms. These are illustrative buffer durations, not the complete round-trip delay.

The complete path can include an input buffer, output buffer, conversion, drivers and plug-ins. An app’s reported round-trip figure may differ from a single-buffer calculation, and reported values are not always exact. At the same buffer size, a higher sample rate reduces the time represented by that buffer, but increases processing and storage demands. Do not change a project’s sample rate solely to hit a number without checking compatibility and workload.

How to reduce audio latency when recording

1. Identify what is late. Listen with wired headphones and compare the live monitor with a short saved recording. A delayed monitor, an out-of-sync saved track and a remote call delay require different checks.
2. Select the correct interface and driver. In the recording app’s audio settings, confirm the intended input and output. On Windows, use a compatible manufacturer-recommended low-latency driver where available; settings vary by device.
3. Lower the buffer gradually. Try the next smaller available setting and repeat the same phrase or instrument part. Stop lowering it when clicks, pops or dropouts appear; move back to a stable setting.
4. Check plug-ins. Bypass effects that add delay during the test. Some apps offer a low-latency recording mode; check what it bypasses before relying on the sound you hear.
5. Try direct monitoring if supported. Route the input through the interface’s direct-monitor control and turn off the app’s monitoring of that same input if you hear it twice. Follow the specific device instructions.
6. Record and play back a test. Confirm the correct input was captured, the performance is aligned as expected, and the file contains no glitches. Raise the buffer again for a demanding mix if real-time monitoring is no longer needed.

In Logic Pro, Apple documents Logic Pro → Settings → Audio → Devices → I/O Buffer Size and a separate Record → Low Latency Mode. Labels may vary by version. Apple recommends increasing the buffer if a smaller setting causes overloads; Ableton likewise explains the buffer and processor tradeoff.

Direct monitoring vs. software monitoring

With direct monitoring, a compatible interface routes its input to headphones or speakers without the usual trip through the recording software. That can make the performer’s monitor much more immediate. Focusrite’s direct-monitoring explanation notes that monitoring the same input through software at the same time can produce two versions of the signal.

Direct monitoring usually does not include effects running in the DAW, so a singer may hear a dry voice even when a plug-in is active in the project. Software monitoring lets you hear the app’s processing but includes its processing path. The available controls and routing depend on the interface and app. See microphone monitoring explained for headset sidetone and doubled-audio checks.

Is monitoring delay the same as a recording that is out of sync?

No. Monitoring delay is what you hear while performing. A recorded timing offset is a mismatch in the saved tracks. An app may compensate for reported device delay when placing a recording, but direct monitoring does not by itself prove the saved track is aligned. Record a short reference, inspect playback against it, and consult your app’s timing or driver-compensation instructions if the offset persists.

For a video with sound already out of sync, first determine whether the audio is consistently early or late throughout the clip. A fixed offset and a drift that grows over time have different causes. The audio replacement guide covers timeline placement and final sync checks.

Common questions

Is a 128-sample buffer always best?

No. It can feel responsive on one system and cause dropouts on another. Choose a stable setting that is comfortable for the task and test the complete project.

Why do I hear my microphone twice?

Direct monitoring and software monitoring may both be active. Another app or an operating-system listening feature can also create a second path. Change one setting at a time and see the microphone echo checklist.

Can a microphone test measure latency?

The Saturalabs microphone test checks the selected input and plays back a short recording; it does not measure or correct round-trip latency. Use your app’s reported values or a suitable measurement procedure for that task.

Does direct monitoring record the effects I hear?

It depends on the routing. A direct hardware path often bypasses software effects; check the interface and DAW setup and listen to the saved track.